http://ueg.sagepub.com/United European Gastroenterology Journal

http://ueg.sagepub.com/content/1/3/175The online version of this article can be found at:

 DOI: 10.1177/2050640613485238

2013 1: 175 originally published online 10 April 2013United European Gastroenterology JournalHidekazu Suzuki, Juntaro Matsuzaki, Sawako Okada, Kenro Hirata, Seiichiro Fukuhara and Toshifumi Hibi

JapanValidation of the GerdQ questionnaire for the management of gastro-oesophageal reflux disease in

  

Published by:

http://www.sagepublications.com

On behalf of: 

United European Gastroenterology

can be found at:United European Gastroenterology JournalAdditional services and information for    

  http://ueg.sagepub.com/cgi/alertsEmail Alerts:

 

http://ueg.sagepub.com/subscriptionsSubscriptions:  

http://www.sagepub.com/journalsReprints.navReprints:  

http://www.sagepub.com/journalsPermissions.navPermissions:  

What is This? 

- Apr 10, 2013OnlineFirst Version of Record  

- Jun 11, 2013Version of Record >>

by guest on October 11, 2013ueg.sagepub.comDownloaded from by guest on October 11, 2013ueg.sagepub.comDownloaded from by guest on October 11, 2013ueg.sagepub.comDownloaded from by guest on October 11, 2013ueg.sagepub.comDownloaded from by guest on October 11, 2013ueg.sagepub.comDownloaded from by guest on October 11, 2013ueg.sagepub.comDownloaded from by guest on October 11, 2013ueg.sagepub.comDownloaded from by guest on October 11, 2013ueg.sagepub.comDownloaded from by guest on October 11, 2013ueg.sagepub.comDownloaded from by guest on October 11, 2013ueg.sagepub.comDownloaded from

Original Article

Validation of the GerdQ questionnaire forthe management of gastro-oesophagealreflux disease in Japan

Hidekazu Suzuki, Juntaro Matsuzaki, Sawako Okada, Kenro Hirata,Seiichiro Fukuhara and Toshifumi Hibi

AbstractBackground: The GerdQ scoring system may be a useful tool for managing gastro-oesophageal reflux disease. However,

GerdQ has not been fully validated in Asian countries.

Objective: To validate the Japanese version of GerdQ and to compare this version to the Carlsson-Dent questionnaire (CDQ)

in both general and hospital-based populations.

Methods: The questionnaires, including the Japanese versions of GerdQ and CDQ, and questions designed to collect

demographic information, were sent to a general population via the web, and to a hospital-based population via conven-

tional mail. The optimal cutoff GerdQ score and the differences in the characteristics between GerdQ and CDQ were assessed.

Results: The answers from 863 web-responders and 303 conventional-mail responders were analysed. When a GerdQ cutoff

score was set at 8, GerdQ significantly predicted the presence of reflux oesophagitis. Although the GerdQ scores were

correlated with the CDQ scores, the concordance rates were poor. Multivariate analysis results indicated that, the additional

use of over-the-counter medications was associated with GerdQ score� 8, but not with CDQ score� 6.

Conclusions: The GerdQ cutoff score of 8 was appropriate for the Japanese population. Compared with CDQ, GerdQ was more

useful for evaluating treatment efficacy and detecting patients’ unmet medical needs.

KeywordsCutoff, GerdQ, GERD, questionnaire, unmet medical need

Received: 22 January 2013; accepted: 12 March 2013

Introduction

The prevalence of gastro-oesophageal reflux disease(GERD) is dramatically increasing in both Westernand Asian countries, including Japan.1 To diagnoseGERD objectively, invasive examinations such as oeso-phagogastroduodenoscopy and pH monitoring can beemployed. However, these methods are inconvenient topatients and have limited availability for primary carephysicians. Therefore, the current guidelines recom-mend a symptom-based approach for diagnosis andtreatment, especially in the primary care of youngpatients, who have a short disease history and noalarm symptoms.2 Primary care physicians experiencethe challenge of accurately diagnosing and effectivelymanaging GERD with drugs that meet patients’satisfaction.

GerdQ is a self-administered 6-item questionnairethat was recently developed as a tool to improve and

standardize symptom-based diagnosis and evaluationof treatment response in patients with GERD.3

Norwegian researchers assessed the diagnostic validityof GerdQ, and they concluded that GerdQ is a useful,complementary tool for diagnosing GERD in primarycare.4 They also reported that a symptom-basedapproach using GerdQ reduced healthcare costs with-out a loss in efficacy.5

In the development of GerdQ, GERD was diag-nosed if patients fulfilled at least one of the followingcriteria: (i) oesophageal pH< 4 for >5.5% of a 24-hperiod, (ii) Los Angeles (LA) grade A�D oesophagitisat endoscopy, (iii) indeterminate 24-h oesophageal pH

Keio University School of Medicine, Tokyo, Japan

Corresponding author:Hidekazu Suzuki, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan.

Email: hsuzuki@a6.keio.jp

United European Gastroenterology Journal

1(3) 175–183

! Author(s) 2013

Reprints and permissions:

sagepub.co.uk/journalsPermissions.nav

DOI: 10.1177/2050640613485238

ueg.sagepub.com

in combination with a positive response to 14 days’esomeprazole treatment, and (iv) positive (>95) symp-tom association probability (SAP).3 This suggests thatthe GerdQ scores might be well correlated with theseverity of acid reflux. In the East Asia and Pacificregions, lower gastric acid secretion is generallyobserved, and the prevalence of reflux oesophagitis iscomparatively low.6–8 Therefore, the validation ofGerdQ in Asian countries is important.

In the present study, we evaluated the usefulness ofthe Japanese version of GerdQ in two differentJapanese populations: general and hospital-basedpopulations. Among the hospital-based population,the association between the GerdQ score and the pres-ence of reflux oesophagitis was assessed. In addition, toclarify the characteristics of GerdQ, the differences inthe characteristics between the GerdQ and theCarlsson-Dent questionnaire (CDQ), which is a trad-itional questionnaire used for diagnosing GERD,9–11

were investigated.

Materials and methods

Study population

The present study was approved by the ethics commit-tee of the Keio University School of Medicine(2010–319, 5 April 2011). The questionnaires, whichcomprised the Japanese versions of GerdQ and CDQ,demographic information (age, gender, smoking habits,alcohol consumption, height, weight, and past medicalhistory), and treatment history for upper gastrointes-tinal symptoms involving prescription or over-the-counter (OTC) medicine use, were sent to both aweb-based and conventional mail-surveyed population.Participants for the web-based survey were randomlyselected from subjects who had registered with a web-survey company and who had a present or past historyof heartburn and/or regurgitation. They providedinformed consent by checking a web box. On theother hand, patients who underwent oesophagogastro-duodenoscopy at the Keio University Hospital betweenDecember 2010 and May 2011 were enrolled in the mailsurvey. The mails were sent in June 2011. Participantswith malignant diseases, peptic ulcers, a history of gas-tric surgery, or systematic diseases affecting the uppergastrointestinal tract were excluded. Participants withheartburn and/or regurgitation more frequently thanonce a week were defined as having GERD.12 The pres-ence of erosive oesophagitis or positive proton pumpinhibitor (PPI) response has not been taken intoaccount for the diagnosis of GERD in the presentstudy.

Individuals were categorized as ‘non-smokers’, ‘ex-smokers’, ‘1–15 cigarettes/day’, and ‘> 15 cigarettes/

day’, according to the number of cigarettes consumedper day. With regard to alcohol consumption, individ-uals were categorized as ‘abstainers’, ‘social drinkers’,‘stopped drinking’, ‘1–2 days/week’, ‘3–4 days/week’,and ‘5–7 days/week’, according to the number of daysalcohol was consumed per week. Body mass index(BMI, weight/height2) was calculated. Participantswere diagnosed with metabolic syndrome if they wereoverweight (BMI� 25 kg/m2) or they had hypertension,diabetes mellitus, or dyslipidaemia, according to theirresponses on the questionnaire.13 For participants whoresponded via the mail survey, the severity of refluxoesophagitis was investigated using medical records.

Statistical analysis

Comparisons of two categorical variables such asgender and the presence/absence of diabetes, hyperten-sion, dyslipidaemia, fatty liver, or the use of each medi-cation were analysed using Fisher’s Exact test. Othercategorical variables such as smoking and alcohol con-sumption were analysed using Pearson’s �2 test.Continuous variables such as age and BMI were ana-lysed using Student’s t-test. The cutoff value for theCDQ was set at 6, whereas the cutoff value of GerdQwas set at 8, in accordance with previous reports.3,9 TheGerdQ cutoff value for predicting the presence of refluxoesophagitis was validated among the mail-surveyedpopulation by using receiver operating characteristic(ROC) analysis. Subsequently, the correlations betweenthe CDQ and GerdQ scores were analysed usingPearson’s correlation analysis, and the concordancerates (kappa coefficient) were calculated betweenCDQ score� 6 and GerdQ score� 8. Finally, the asso-ciations of CDQ score� 6 or GerdQ score� 8 withdemographic factors were analysed using univariateand multivariate logistic regression models. The multi-variate logistic regression model was adjusted for age,gender, the presence of metabolic syndrome, and theuse of both prescription and OTC medications. All stat-istical analyses were conducted using SPSS Statisticsversion 18.0 for Windows software (SPSS Japan,Tokyo, Japan). Two-sided p-values were consideredstatistically significant at a level of 0.05.

Results

Characteristics of participants

In the web survey, a total of 1630 participants were sentquestionnaires, among which 1024 complete responseswere received (Figure 1a). After excluding participantswith malignant diseases (n¼ 30) and those with pepticulcers (n¼ 131), the responses of the remaining 863 par-ticipants were used in the analysis. Among them, 362

176 United European Gastroenterology Journal 1(3)

participants were included in the GERD group. Fromthe conventional mail survey, a total of 1109 patientswere enrolled (Figure 1b). After excluding participantswith peptic ulcers (n¼ 32), histories of gastric surgery(n¼ 7), malignant diseases (n¼ 55), severe systematicdiseases (n¼ 22), and unknown addresses (n¼ 153),840 patients were sent questionnaires along with aninformed consent form; of these, 303 patients gavecomplete responses. Among the 303, 127 participantswere included in the GERD group. Demographic char-acteristics of these two populations were totally differ-ent except for BMI, as shown in Table 1.

Optimal cutoff value of GerdQ for the prediction ofreflux oesophagitis

Among the 303 included mail respondents, 21 patientshad LA grade A oesophagitis, 12 had LA grade Boesophagitis, one had LA grade C oesophagitis, onehad LA grade D oesophagitis, and the remaining 268patients did not have reflux oesophagitis. To validatethe optimal cutoff value of GerdQ for predicting the

presence of reflux oesophagitis, the sensitivity, specifi-city, positive predictive values (PPVs), negative predict-ive values (NPVs), and area under the ROC curves(AUCs) were calculated (Table 2). When the cutoffvalue was set at 8, the PPV and AUC were the highest.A significant association between GerdQ-positive andthe presence of reflux oesophagitis was also shown(p¼ 0.02). These results revealed that a cutoff value of8 was also most predictive of reflux oesophagitis inJapanese population. In addition, GerdQ score� 8showed higher specificity, PPV, and AUC than CDQscore� 6 (Table 2). These data indicated that GerdQ ismore useful than CDQ for predicting the presence ofreflux oesophagitis, although the sensitivity of GerdQscore� 8 was low (34.3%).

Differences in characteristics between CDQscore� 6 and GerdQ score� 8

The GerdQ and CDQ scores were significantly corre-lated in both the web-based and mail-surveyed popula-tion (Figure 2). However, many participants showed

Web-based population Mail-surveyed population

Participants who weresent questionnaires

via web siten=1,630

Complete thequestionnaire

n=1,024

Whole analyzedpopulation

n=863

excluded

Malignancy (n=30)Peptic ulcer (n=131)

Freqency ofreflux symptom≥ once a week

No

Yes

GERDn = 362 (41.9%)

Patients whounderwent EGD

n=1,109

Sent questionnairesvia conventional mail

n=840

Whole analyzedpopulation

n=303

Patients with no answer(n = 537)

excluded

Yes

NoFreqency ofreflux symptom≥ once a week

GERDn = 127 (41.9%)

Non-GERDn = 501(58.1%)

Non-GERDn = 176(58.1%)

Peptic ulcer (n=32)History of gastric surgery(n=7)Malignancy (n=55)Systematic disease(n=22)Address unknown (n=153)

Figure 1. The web-based (a) and mail-surveyed (b) populations.

EGD, oesophagogastroduodenoscopy; GERD, gastro-oesophageal reflux disease.

Suzuki et al. 177

CDQ score� 6 and GerdQ score< 8, or CDQ score< 6and GerdQ score� 8, even in the GERD groups. Thissuggests that there are differences in the patient char-acteristics between populations of patients with CDQ

score� 6 and GerdQ score� 8 populations. Since CDQcutoff score is sometimes set at 4 in Japan, kappa coef-ficients between CDQ score� 4 and GerdQ score� 8were also calculated. Using CDQ cutoff score of 6,

Table 1. Characteristics of analysed patients

Web-based population Mail-survey population Web vs. mail

Whole

(n¼ 863)

Non-GERD

(n¼ 501)

GERD

(n¼ 362) p-value

Whole

(n¼ 303)

Non-GERD

(n¼ 176)

GERD

(n¼ 127) p-value p-value

Age (years) 41.1� 9.3 40.4� 9.4 41.9� 9.1 0.02a 50.3� 15.4 51.2� 14.4 49.1� 16.3 0.02a <0.001a

Gender

Men 391 (45.3) 218 (43.5) 173 (47.8) 0.24b 168 (55.4) 100 (56.8) 68 (53.5) 0.64b 0.003b

Women 472 (54.7) 283 (56.5) 189 (52.2) 135 (44.6) 76 (43.2) 59 (46.5)

Smoking habits

Non-smokers 409 (47.4) 246 (49.1) 163 (45.0) 0.58c 180 (59.4) 112(64.0) 68(53.5) 0.13c <0.001c

Ex-smokers 253 (29.3) 146 (29.1) 107 (29.6) 88 (29.1) 47(26.9) 41(32.3)

1–15/day 104 (12.1) 57 (11.4) 47 (13.0) 11 (3.6) 7(4.0) 4(3.1)

> 15/day 97 (11.2) 52 (10.4) 45 (12.4) 23 (7.6) 9(4.6) 14(11.0)

Alcohol habits

Abstainers 161 (18.7) 114 (22.8) 47 (13.0) 0.004c 61 (20.1) 29(16.6) 32(25.2) 0.22c <0.001c

Social drinkers 329 (38.1) 182 (36.3) 147 (40.6) 89 (29.4) 48(27.4) 41(32.3)

Stop drinking 43 (5.0) 28 (5.6) 15 (4.1) 16 (5.3) 9(5.1) 7(5.5)

1–2 days/week 58 (6.7) 36 (7.2) 22 (6.1) 46 (15.2) 31(17.7) 15(11.8)

3–4 days/week 76 (8.8) 40 (8.0) 36 (9.9) 38 (12.5) 26(14.9) 12(9.4)

5–7 days/week 196 (22.7) 101 (20.2) 95 (26.2) 52 (17.2) 32(18.3) 20(15.7)

BMI (kg/m2) 23.2� 4.3 23.0� 4.3 23.4� 4.3 0.26a 22.9� 7.8 22.9� 7.3 22.9� 8.3 0.99a 0.27a

Diabetes mellitus 33 (3.8) 13 (2.6) 20 (5.5) 0.03b 27 (8.9) 16(9.1) 11(8.7) 1.00b 0.001b

Hypertension 101 (11.7) 53 (10.6) 48 (13.3) 0.23b 55 (18.2) 32(18.1) 23(18.1) 1.00b 0.01b

Dyslipidaemia 75 (8.7) 32 (6.4) 43 (11.9) 0.005b 52 (17.2) 32(18.1) 20(15.7) 0.65b <0.001b

Fatty liver 63 (7.3) 34 (6.8) 29 (8.0) 0.50b 37 (12.2) 16(9.1) 21(16.5) 0.07b 0.01b

Medications for upper gastrointestinal tract

PPI 42 (4.9) 20 (4.0) 22 (6.1) 0.16b 73 (24.1) 35 (20.2) 38 (30.6) 0.04b <0.001b

H2RA 44 (5.1) 17 (3.4) 27 (7.5) 0.007b 34 (11.2) 16 (9.2) 18 (14.5) 0.20b <0.001b

Prokinetics 24 (2.8) 10 (2.0) 14 (3.9) 0.10b 30 (9.9) 9 (5.2) 21 (17.1) 0.001b <0.001b

Others 35 (4.1) 16 (3.2) 19 (5.2) 0.08b 35 (11.6) 12 (7.0) 23 (18.5) 0.003b <0.001b

OTC medications 175 (20.3) 55 (11.0) 120 (33.1) <0.001b 30 (9.9) 5 (2.8) 25 (19.7) <0.001b <0.001b

Values are mean� standard deviation or n (%). aStudent’s t-test; bFisher’s Exact test; cPearson’s �2 test.

BMI, body mass index; H2RA, histamine H2-receptor antagonist; PPI, proton pump inhibitor; OTC, over-the-counter.

Table 2. Sensitivity and specificity of CDQ and GerdQ for the diagnosis of reflux oesophagitis among the mail-survey population (n¼ 303)

Sensitivity (%) Specificity (%) PPV (%) NPV (%) AUC p-value

CDQ� 6 51.4 61.9 15.0 90.7 0.567 0.14

GerdQ� 6 88.6 14.6 11.9 90.7 0.516 0.80

GerdQ� 7 42.9 70.5 16.0 90.4 0.567 0.12

GerdQ� 8 34.3 82.5 20.3 90.6 0.584 0.02

GerdQ� 9 17.1 89.6 17.6 89.2 0.533 0.25

p-values calculated using Fisher’s Exact test.

CDQ, Carlsson-Dent questionnaire; PPV, positive predictive value; NPV, negative predictive value; AUC, area under the curve.

178 United European Gastroenterology Journal 1(3)

kappa coefficients were 0.24 in web-based populationand 0.27 in mail-surveyed population, whereas, usingCDQ cutoff score of 4, kappa coefficients were 0.22 inweb-based population and 0.25 in mail-surveyed popu-lation. All of the kappa coefficients were< 0.4, whichindicates poor concordance.

Therefore, the average score of each question ofGerdQ was compared between participants with CDQscore� 6 and CDQ score< 6 (Table 3). The scores ofquestion 1, 2, 5, and 6 in GerdQ were significantly

higher in CDQ score� 6 than CDQ score< 6. On theother hand, the score of question 3 in GerdQ was sig-nificantly lower in CDQ score� 6 than CDQ score< 6.The score of question 4 in GerdQ was not differentbetween CDQ score� 6 and CDQ score< 6. Theseshowed that question 3 and 4 in GerdQ caused poorconcordance about the diagnosis of GERD usingGerdQ and CDQ.

Subsequently, the associations of CDQ score� 6and GerdQ score� 8 with demographic information

18

16

14

12

10

8

6

4

2

0

18

16

14

12

10

8

6

4

2

0

18

16

14

12

10

8

6

4

2

0

18

16

14

12

10

8

6

4

2

0

-2 0 2 4 6 8 10 12 14 16

-2 0 2 4 6 8 10 12 14 16

-2 0 2 4 6 8 10 12 14 16

-2 0 2 4 6 8 10 12 14 16CDQ

CDQCDQ

CDQ

Ger

dQ

Ger

dQ

Ger

dQ

Ger

dQ

Whole analyzed population Whole analyzed population

GERD population GERD population

R = 0.126p < 0.001k = 0.10

R = 0.222p < 0.05k = 0.23

R = 0.374p < 0.001k = 0.37

R = 0.346p < 0.001k = 0.24

Number of participants in a dot

1 - 9 10 - 19 20 - 29 30 - 39 40 -

(A) (C)

(D)(B)

Web-based questionnaire Mail-surveyed questionnaire

Figure 2. Correlation and concordance between CDQ and GerdQ scores in the total analysed population and in the GERD groups.

CDQ, Carlsson-Dent questionnaire; GERD, gastro-oesophageal reflux disease.

Suzuki et al. 179

were analysed among GERD groups using logisticregression models. The results of univariate analysisinvolving GERD patients responding to the web-based survey showed that, men and the presenceof metabolic syndrome were associated with CDQscore� 6, whereas older age, the presence of meta-bolic syndrome, and the use of both prescriptionand OTC medications were associated with GerdQscore� 8 (Table 4). Among GERD patients in themail-survey population, no association was observedbetween CDQ score� 6 and demographic factors,whereas men and the presence of metabolic syn-drome were associated with GerdQ score� 8. Theuse of both prescription and OTC medications wasmarginally associated with a GerdQ score� 8(p¼ 0.06).

Finally, multivariate logistic regression analysis wasperformed with adjustments for age, gender, the pres-ence of metabolic syndrome, and the use of both pre-scription and OTC medications (Table 5). According tothe multivariate analysis results, among GERDpatients in the web-surveyed population, men wereassociated with CDQ score� 6, whereas older age andthe use of both prescription and OTC medications wereindependently associated with GerdQ score� 8.Contrarily, among GERD patients in the mail-surveyedpopulation, men and the use of both prescription andOTC medications were independently associated withGerdQ score� 8. These results suggest that the use ofboth prescription and OTC medications would be

associated with GerdQ score� 8, but not CDQscore� 6, in any population.

Discussion

When the GerdQ was developed as an exploratory partof the DIAMOND study, a cutoff of 8 showed thehighest specificity (71.4) and sensitivity (64.6) forGERD. The present study also showed that a GerdQcutoff of 8 gave the best balance with regard to sensi-tivity and specificity for reflux oesophagitis in Japanesepopulations. The reason for the low positive predictivevalue (20.3) would be that the prevalence of refluxoesophagitis is low in GERD patients. Previous datashowed that the prevalence of non-erosive reflux dis-ease (NERD) in medical check-up studies was about70�80% in Asian GERD population.14 Patients withGerdQ score� 8 but without reflux oesophagitis werethought to be NERD patients.

According to the univariate logistic regression ana-lysis, the presence of metabolic syndrome was asso-ciated with GerdQ score� 8 in both the web- andmail-surveyed populations. Because metabolic syn-drome is well known to be associated with the develop-ment and progression of GERD,15 the associationbetween GerdQ score� 8 and metabolic syndromealso suggests that GerdQ is useful for diagnosingGERD in Japanese individuals. On the other hand,CDQ score� 6 was also associated with the presenceof metabolic syndrome in the web-surveyed population,

Table 3. Scores for each question of GerdQ in CDQ-positive and -negative participants

Web-based population Mail-survey population

CDQ< 6

(n¼ 135)

CDQ� 6

(n¼ 227) p-value

CDQ< 6

(n¼ 40)

CDQ� 6

(n¼ 87) p-value

GerdQ

1. How often did you have a burning feeling

behind your breastbone (heartburn)?

0.38� 0.71 0.96� 0.89 <0.001 0.14� 0.50 0.84� 0.93 <0.001

2. How often did you have stomach contents

(liquid or food) moving upwards to your throat

or mouth (regurgitation)?

0.36� 0.63 0.68� 0.80 <0.001 0.26� 0.64 0.83� 0.92 <0.001

3. How often did you have a pain in the centre of

the upper stomach?

2.65� 0.70 2.42� 0.83 <0.001 2.66� 0.74 2.29� 0.91 <0.001

4. How often did you have nausea? 2.57� 0.74 2.53� 0.70 0.37 2.77� 0.66 2.74� 0.69 0.72

5. How often did you have difficulty getting a

good night’s sleep because of your heartburn

and / or regurgitation?

0.18� 0.51 0.32� 0.63 <0.001 0.06� 0.28 0.29� 0.63 <0.001

6. How often did you take additional medication

for your heartburn and / or regurgitation,

other than what the physician told you to take?

0.23� 0.65 0.49� 0.84 <0.001 0.08� 0.39 0.33� 0.77 <0.001

Values are mean� standard deviation. p-values calculated using Student’s t-test.

CDQ, Carlsson-Dent questionnaire.

180 United European Gastroenterology Journal 1(3)

but not in mail-surveyed population. CDQ cannot beused to distinguish between GERD and functionalheartburn,16 and functional heartburn is not associatedwith metabolic syndrome.17 CDQ score� 6 was notassociated with metabolic syndrome in the mail-surveyed population because more patients with func-tional heartburn were included in the population. Inaddition, epigastric pain (question 3 in GerdQ) wasmore frequent in participants with CDQ score� 6than those with CDQ score< 6, suggesting that partici-pants with CDQ score� 6 were likely to have functionaldyspepsia. These results suggest that GerdQ would bemore useful for distinguishing between GERD andfunctional upper gastrointestinal disorders (heartburnand dyspepsia) than CDQ. However, GerdQ has limita-tions to distinguish between GERD and functionalheartburn. Although positive PPI response and SAPwere taken into account to diagnose GERD in thedevelopment of GerdQ, some researchers have reportedthat positive PPI response cause overestimation of func-tional heartburn patients.18,19 SAP might be overinter-preted in patients with refractory GERD especiallywhen low reflux rates are observed.20 In addition, non-acid reflux, which is reported to be involved in the devel-opment of reflux symptoms, was not taken into accountin GerdQ.21 Therefore, to validate how useful for dis-tinguishing between GERD and functional heartburn,GerdQ should be evaluated using combined impedancepH monitoring.

According to the multivariate logistic regressionanalysis, the additional use of OTC medicationswas well associated with GerdQ score� 8 in bothpopulations. The additional use of OTC medicationsmay indicate that prescription medicines are inad-equate for treating reflux symptoms. This result sug-gests that the GerdQ would be useful for evaluatingtreatment response and for detecting unmet medicalneeds.

Limitations of the present study include the absenceof pH monitoring for the mail-surveyed population.The pH monitoring data might have revealed the preva-lence of NERD and functional heartburn in this popu-lation, and the optimal cutoff value of GerdQ and thedifferences between GerdQ and CDQ might have beenmore clearly evaluated. The low response rate for thequestionnaires in the mail-surveyed population mighthave caused selection bias.

In conclusion, the present study is the first study toevaluate the usefulness of GerdQ in Japanese popula-tion. The GerdQ cutoff score of 8 was found to also beappropriate for the Japanese population. The GerdQwas better able to detect unmet therapeutic needs.Symptom-based management of GERD using GerdQwould be beneficial in Japan, as was revealed inWestern countries.Ta

ble

4.A

ssoci

ati

on

bet

wee

nth

esc

ore

sof

CD

Qan

dG

erd

Qan

dd

emogra

ph

icfa

ctors

am

on

gG

ER

Dp

ati

ents

(un

ivari

ate

an

aly

sis)

Web

-base

dp

op

ula

tion

Mail

-su

rvey

pop

ula

tion

CD

Q<

6

(n¼

135)

CD

Q�

6

(n¼

227)

OR

(95%

CI)

Ger

dQ<

8

(n¼

183)

Ger

dQ�

8

(n¼

179)

OR

(95%

CI)

CD

Q<

6

(n¼

40)

CD

Q�

6

(n¼

87)

OR

(95%

CI)

Ger

dQ<

8

(n¼

71)

Ger

dQ�

8

(n¼

87)

OR

(95%

CI)

Age

(yea

rs)

40.8�

9.2

42.6�

9.0

1.02

(1.0

0–1.

05)

40.2�

9.1

43.7�

8.8

1.05

(1.0

2–1.

07)

50.3�

6.0

48.5�

9.0

0.97

(0.9

3–1.

02)

49.1�

7.9

49.0�

8.6

1.00

(0.9

6–1.

04)

Gen

der

(men

)52

(38.

5)12

1(5

3.3)

1.82

(1.1

8–2.

81)

88(4

8.1)

85(4

7.5)

0.98

(0.6

5–1.

48)

22(5

5.0)

46(5

2.9)

0.92

(0.4

3–1.

95)

29(4

0.8)

39(6

9.6)

3.32

(1.5

8–6.

97)

Hea

vysm

oki

ng

(>15

/day

)

17(1

2.6)

28(1

2.3)

0.98

(0.5

1–1.

86)

20(4

4.4)

25(5

5.6)

1.32

(0.7

1–2.

48)

2(5

.0)

12(1

3.8)

3.04

(0.6

5–14

.3)

7(9

.9)

7(1

2.5)

1.31

(0.4

3–3.

97)

Hea

vyalc

oh

ol

con

sum

pti

on

(5–7

day

s/w

eek)

28(2

0.7)

67(2

9.5)

1.60

(0.9

7–2.

65)

46(2

5.1)

49(2

7.4)

1.12

(0.7

0–1.

79)

6(1

5.0)

14(1

6.1)

1.09

(0.3

8–3.

07)

11(1

5.5)

9(1

6.1)

1.04

(0.4

0–2.

73)

Met

ab

oli

csy

nd

rom

e49

(36.

3)10

9(4

8.0)

1.62

(1.0

5–2.

51)

68(3

7.2)

90(5

0.3)

1.71

(1.1

3–2.

60)

20(5

0.0)

35(4

0.2)

0.67

(0.3

2–1.

43)

24(3

3.8)

31(5

5.4)

2.43

(1.1

8–4.

99)

Usi

ng

both

pre

scri

pti

on

an

d

OTC

med

icati

on

s

8(5

.9)

26(1

1.5)

2.05

(0.9

0–4.

68)

6(3

.3)

28(1

5.6)

5.47

(2.2

1–13

.6)

4(1

0.0)

7(8

.0)

0.79

(0.2

2–2.

86)

3(4

.2)

8(1

4.3)

3.78

(0.9

5–15

.0)

Valu

esare

mea

n�

SD

or

n(%

),u

nle

ssoth

erw

ise

state

d.

Bold

ind

icate

ssi

gn

ific

an

td

iffe

ren

ces.

CD

Q,

Carl

sson

-Den

tq

ues

tion

nair

e;O

TC,

ove

r-th

e-co

un

ter.

Suzuki et al. 181

Funding

This study was supported in part by the Smoking ResearchFoundation (to HS), the Keio Gijuku Academic

Development Fund (No. 24-24, to HS), a Keio UniversityGrant-in-Aid for Encouragement of Young MedicalScientists (to JM), and a grant from AstraZeneca K.K.

Conflict of interest

The authors declare that there is no conflict of interest.

References

1. Fujiwara Y and Arakawa T. Epidemiology and clinical

characteristics of GERD in the Japanese population.

J Gastroenterol 2009; 44: 518–534.

2. Kahrilas PJ, Shaheen NJ, Vaezi MF, et al. American

Gastroenterological Association Medical Position

Statement on the management of gastroesophageal reflux

disease. Gastroenterology 2008; 135: 1383–1391.3. Jones R, Junghard O, Dent J, et al. Development of the

GerdQ, a tool for the diagnosis and management of

gastro-oesophageal reflux disease in primary care.

Aliment Pharmacol Ther 2009; 30: 1030–1038.

4. Jonasson C, Wernersson B, Hoff DA, et al. Validation of

the GerdQ questionnaire for the diagnosis of gastro-oeso-

phageal reflux disease. Aliment Pharmacol Ther2013.5. Jonasson C, Moum B, Bang C, et al. Randomised clinical

trial: a comparison between a GerdQ-based algorithm and

an endoscopy-based approach for the diagnosis and initial

treatment of GERD. Aliment Pharmacol Ther 2012; 35:

1290–1300.

6. Wong BC and Kinoshita Y. Systematic review on epidemi-

ology of gastroesophageal reflux disease in Asia. Clin

Gastroenterol Hepatol 2006; 4: 398–407.7. Kinoshita Y, Kawanami C, Kishi K, et al. Helicobacter

pylori independent chronological change in gastric acid

secretion in the Japanese. Gut 1997; 41: 452–458.

8. Kinoshita Y, Adachi K, Hongo M, et al. Systematic review

of the epidemiology of gastroesophageal reflux disease in

Japan. J Gastroenterol 2011; 46: 1092–1103.

9. Carlsson R, Dent J, Bolling-Sternevald E, et al. The use-

fulness of a structured questionnaire in the assessment of

symptomatic gastroesophageal reflux disease. Scand J

Gastroenterol 1998; 33: 1023–1029.10. Mishima I, Adachi K, Arima N, et al. Prevalence of

endoscopically negative and positive gastroesophageal

reflux disease in the Japanese. Scand J Gastroenterol

2005; 40: 1005–1009.11. Iwasaki E, Suzuki H, Sugino Y, et al. Decreased levels of

adiponectin in obese patients with gastroesophageal

reflux evaluated by videoesophagography: possible rela-

tionship between gastroesophageal reflux and metabolic

syndrome. J Gastroenterol Hepatol 2008; 23(Suppl 2):

S216–S221.12. Vakil N, van Zanten SV, Kahrilas P, et al. The Montreal

definition and classification of gastroesophageal reflux

disease: a global evidence-based consensus. Am J

Gastroenterol 2006; 101: 1900–1920.13. Alberti KG, Zimmet P and Shaw J. The metabolic syn-

drome � a new worldwide definition. Lancet 2005; 366:

1059–1062.14. Jung HK. Epidemiology of gastroesophageal reflux dis-

ease in Asia: a systematic review. J Neurogastroenterol

Motil 2011; 17: 14–27.15. Lee YC, Yen AM, Tai JJ, et al. The effect of metabolic

risk factors on the natural course of gastro-oesophageal

reflux disease. Gut 2009; 58: 174–181.

16. Netinatsunton N, Attasaranya S, Ovartlarnporn B, et al.

The value of Carlsson-dent questionnaire in diagnosis

of gastroesophageal reflux disease in area with low

prevalence of gastroesophageal reflux disease.

J Neurogastroenterol Motil 2011; 17: 164–168.

17. Matsuzaki J, Suzuki H, Iwasaki E, et al. Serum lipid

levels are positively associated with non-erosive reflux

disease, but not with functional heartburn.

Neurogastroenterol Motil 2010; 22: 965–970, e251.18. Zerbib F, Belhocine K, Simon M, et al. Clinical, but not

oesophageal pH-impedance, profiles predict response to

proton pump inhibitors in gastro-oesophageal reflux dis-

ease. Gut 2012; 61: 501–506.

Table 5. Association between the scores of CDQ and GerdQ and demographic factors among GERD patients (multivariate analysis)

OR (95% CI)

Web-based population Mail-survey population

CDQ GerdQ CDQ GerdQ

Age 1.01 (0.99–1.04) 1.04 (1.02–1.07) 0.98 (0.93–1.03) 0.96 (0.92–1.02)

Gender (men) 1.66 (1.05–2.61) 0.78 (0.50–1.23) 1.12 (0.49–2.54) 3.66 (1.57–8.53)

Metabolic syndrome 1.29 (0.80–2.07) 1.38 (0.86–2.20) 0.73 (0.32–1.68) 1.88 (0.82–4.28)

Using both prescription and OTC medications 1.96 (0.85–4.52) 5.11 (2.04–12.8) 0.87 (0.23–3.26) 4.61 (1.06–20.2)

Values are odds ratio (95% CI). Bold indicates significant differences.

CDQ, Carlsson-Dent questionnaire; OTC, over-the-counter.

182 United European Gastroenterology Journal 1(3)

19. Savarino E, Marabotto E, Zentilin P, et al. Theadded value of impedance-pH monitoring to Rome IIIcriteria in distinguishing functional heartburn from

non-erosive reflux disease. Dig Liver Dis 2011; 43:542–547.

20. Slaughter JC, Goutte M, Rymer JA, et al. Caution aboutoverinterpretation of symptom indexes in reflux

monitoring for refractory gastroesophageal reflux dis-ease. Clin Gastroenterol Hepatol 2011; 9: 868–874.

21. Savarino E, Tutuian R, Zentilin P, et al. Characteristics

of reflux episodes and symptom association in patientswith erosive esophagitis and nonerosive reflux disease:study using combined impedance-pH off therapy. Am JGastroenterol 2010; 105: 1053–1061.

Suzuki et al. 183